Intermittent CI failures building distributions

[douglascamata]

We are having some intermittent CI failures in PRs when building distributions.

They look like this:

Run make ci
make ci
shell: /usr/bin/bash -e {0}
env:
GOTOOLCHAIN: local
DISTRIBUTIONS: otelcol-contrib
Artifacts to generate: otelcol-contrib,builder,opampsupervisor
Fetching builder-config.yaml...
Verifying all manifest.yaml files in 'distributions'...
Checking distributions/otelcol/manifest.yaml
Checking distributions/otelcol-k8s/manifest.yaml
Checking distributions/otelcol-contrib/manifest.yaml
./scripts/validate-components.sh: line 73: printf: write error: Broken pipe
./scripts/validate-components.sh: line 73: printf: write error: Broken pipe
./scripts/validate-components.sh: line 73: printf: write error: Broken pipe
Checking distributions/otelcol-otlp/manifest.yaml
Checking distributions/otelcol-ebpf-profiler/manifest.yaml

The following components MUST be listed in
https://github.com/open-telemetry/opentelemetry-collector-contrib/tree/main/cmd/otelcontribcol/builder-config.yaml
to ensure that they can be built:

github.com/open-telemetry/opentelemetry-collector-contrib/exporter/sumologicexporter
github.com/open-telemetry/opentelemetry-collector-contrib/exporter/tencentcloudlogserviceexporter
github.com/open-telemetry/opentelemetry-collector-contrib/extension/awsproxy

make: *** [Makefile:43: validate-components] Error 1

Above you can see a failure to build the Contrib distributions.

This happens even in completely unrelated PRs, like #1209. See the full build log at https://github.com/open-telemetry/opentelemetry-collector-releases/actions/runs/18555999587/job/52893535451?pr=1209 for more details.

[dosubot]

The intermittent CI failures and "Broken pipe" errors are caused by the validate-components.sh script checking that every component listed in your distribution's manifest.yaml files is also present in the upstream builder-config.yaml for Contrib. If a component (like sumologicexporter, tencentcloudlogserviceexporter, or awsproxy) is missing from builder-config.yaml, the script fails and exits with an error, which breaks the CI build—even for unrelated PRs, since the validation runs on every build and always checks the latest upstream config. The "Broken pipe" itself is a side effect of how the script pipes output to grep, but the real issue is the missing component listing, not the pipe error itself.

To resolve these failures, make sure all required components in your distribution manifests are explicitly listed in the upstream builder-config.yaml file. This keeps your manifests and builder config in sync and prevents CI from failing due to missing components. More details on how the script works and why this happens can be found here.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}    

[mowies]

do you have any idea what could cause that? i also saw it a few times in the past already

[mowies]

i don't think it's components that are actually missing

[douglascamata]

I think a good hint for the source of the problem might be here: ./scripts/validate-components.sh: line 73: printf: write error: Broken pipe. It happens inside the contrib distro and it's the distro that fails afterwards.

So I guess somehow the file could be malformed because of some issue in scripts/automation?

[mowies]

happened again here: https://github.com/open-telemetry/opentelemetry-collector-releases/actions/runs/18650203271/job/53166255745?pr=1213

[jackgopack4]

looking into this

[mowies]

seems like there's still some CI issues: https://github.com/open-telemetry/opentelemetry-collector-releases/actions/runs/18748248757/job/53480772151?pr=1154

[douglascamata]

seems like there's still some CI issues: open-telemetry/opentelemetry-collector-releases/actions/runs/18748248757/job/53480772151?pr=1154

Reopening this then.

@jackgopack4: maybe we will also need your changes from #1215 and/or something to help us debug when this happens (simply cat the files, I guess)?